imagesearch
Latest
Google image search results crammed into picture dictionary
Though not quite a replacement for Mountain View's ill-fated dictionary, this 1,240 page tome contains the first Google image search result for each word in a run-of-the-mill dictionary. With a PHP script, London-based artists Felix Heyes and Ben West scraped the image from each search and compiled an alphabetically ordered PDF brimming with 21,000 images -- safesearch-disabled warts and all. "It's really an unfiltered, uncritical record of the state of human culture in 2012," West told Creative Applications Network. Alas, the volume isn't destined for mass distribution -- presumably to avoid copyright issues -- but the pair is considering sending a small batch of soft cover copies to print.
Google Images get spruced up, don't need no stinkin' text
Those alchemists over in Mountain View have been fiddling with their search engine again and the product has been one of the biggest redesigns ye olde Google has received to date. Gone are the little captions and size measurements under each image -- well, not gone, just hidden away until you hover over a pic -- to be replaced by a densely packed compendium of your results, which just keeps going and going. Seriously, the new Images search can fit up to 1,000 pictures on one page, with thumbnails loading in a logical top-to-bottom order. A new landing page has also been implemented, showing you the image you selected superimposed on top of the website it belongs to, making for a more streamlined search experience overall. We like it, it's fast and it's pretty cool, but is anyone working on result relevance at all? Try looking for an iPhone 4 snapshot and you'll have to scroll past 43 mockups before finding the real thing.
Google working with D-Wave on what may or may not be quantum computing
When we first mentioned D-Wave way back in early 2007 we immediately compared it to Steorn -- less than optimal beginnings. The company was promising quantum computing for the masses and, while it did demonstrate a machine that exhibited qubit-like behavior, the company never really silenced critics who believed the underpinnings of the machine were rather more binary in nature. Those disbelievers are surely shutting up now, with word hitting the street that Google has signed on, building new image search algorithms that run on D-Wave's C4 Chimera chip. The first task was to learn to spot automobiles in pictures, something that the quantum machine apparently learned to do simply by looking at other pictures of cars. It all sounds rather neural-networkish to us, but don't let our fuzzy logic cloud your excitement over the prospect of honest to gosh commercial quantum computing.
